About

Senior AI consultant, designing AI-powered finance and supply-chain analytics for large enterprises — from source data to the dashboard a CFO actually trusts. Master’s in Data Science; four years shipping production AI systems before that. I write this series because most finance AI fails before a model is chosen, and almost nobody talks about that part.

What I do now

I work in enterprise AI consulting. My day-to-day is building analytics and agentic AI systems for finance and supply-chain teams: tracing KPIs from source systems to the screen, reconciling numbers that don’t match, defining what an AI system may decide alone and where a human must sign off. Client work is confidential, so on this site I write about the patterns, never the projects.

Before consulting

I spent four years as an AI engineer shipping production systems: real-time computer vision for retail shelf intelligence, neural networks optimised to run on microcontrollers and embedded cameras, RAG pipelines and LLM-powered log analytics. That period taught me the difference between a model that works in a notebook and a system a business can rely on — which is, in the end, what this whole series is about.

Skills

  • AI and GenAI — agentic systems, LLM integration and evaluation, RAG and vector retrieval, prompt engineering, fine-tuning (PEFT)
  • Engineering — Python, FastAPI, LangChain / LangGraph, CI/CD, Docker, Kubernetes
  • Data and analytics — enterprise data integration, KPI lineage and reconciliation, PostgreSQL, vector databases
  • Computer vision and edge AI — object detection, TensorRT and ONNX optimisation, deployment on embedded devices
  • Cloud — Azure, AWS, GCP

Education

  • Master of Technology, Data Science — Amrita School of Engineering, Coimbatore (2022)
  • Bachelor of Technology, Mechanical Engineering — Amrita School of Engineering, Coimbatore (2019)

Why this series

Most AI projects in finance fail before a model is ever chosen — in the unglamorous period where decisions, definitions and controls should have been agreed and weren’t. I’ve watched it happen from inside the systems. Finance × AI Weekly is one practical idea a week about getting that part right.

Find me on LinkedIn or use the contact form.